Widen pseudo for numerical work: sized types, FLOAT32, exact built-ins, USE IEEE

Pseudo could say what a procedure computes only up to INTEGER and FLOAT,
so it could not state what a Mojo or Python program with Int32, UInt8 or
Float32 values will compute, and it had no words for rounding itself.

- lang/parse: INT8..INT64, UINT8..UINT64, FLOAT32, FLOAT64 as types
  (INTEGER or FLOAT with a width, so every "whole number" rule accepts
  them); conversions written as Mojo writes them, INT32(X), FLOAT32(X);
  1_000_000, 0xFF and 0b1010 literals; a type name as an argument
  (MAX_VALUE(INT8)); the USE line (USE IEEE now, library modules next).
- check: widths and precisions never mix silently (mixed-width,
  mixed-precision); a literal takes the type of the sized value beside
  it, as in Mojo, and one that does not fit is an error before any run
  (literal-range); typing for every new built-in.
- interp: a sized result, store, parameter, RETURN or conversion that
  does not fit stops the run (kind overflow), WRAP(X, T) wraps on
  purpose; FLOAT32 results are rounded to single precision; USE IEEE
  lets NaN and infinities flow.
- numeric.js: exact helpers: FMA rounded once, EXACT(X) with every digit
  a FLOAT holds, ULP and NEXT_AFTER for both precisions, bit operations
  in two's complement.
- Built-ins: ASIN ACOS ATAN SINH COSH TANH HYPOT POW CBRT EXPM1 LOG1P E
  TRUNC SIGN COPYSIGN CLAMP FMA IS_NAN IS_INF IS_FINITE INF NAN ULP
  NEXT_AFTER IS_CLOSE EXACT MAX_VALUE MIN_VALUE EPSILON GCD LCM BIT_AND
  BIT_OR BIT_XOR BIT_NOT SHIFT_LEFT SHIFT_RIGHT WRAP. They stay usable
  as names (a variable E, a FUNCTION GCD): the program's own wins.
- vcgen: an overflow obligation on every sized result, then the range
  as a fact (also for sized parameters and loop-cut variables); FLOAT32
  arithmetic as its own uninterpreted functions, constants rounded
  exactly, TOFLOAT32(v) otherwise; argument obligations for SQRT, the
  logarithms, ASIN/ACOS, LOG1P, CLAMP and the shifts; nothing proved
  under USE IEEE.
- testing: the fuzzer tries the edges of a sized type and generates
  FLOAT32 values.
- pseudo-to-mojo: Int8..Float32, conversions, the new built-ins through
  std.math (each name in the harvested index), a note where Mojo wraps.
- Tests: numeric.test.mjs; corpus 49-52; prover fixtures int8-overflow,
  float32-store, ieee-nan; the reference documents every new form.

Gate: 324 pass (core). New Mojo uses std.math names the interpreter has
not been verified on (listed in sync-mojo's KNOWN_GAPS); they need
pixi run mojo-verify on the network before the interpreter runs them.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FgvbRdoEheAHqsRmr2WuKZ
This commit is contained in:
alexmandClaude committed 2026-10-05 11:43:53 +00:00
1 parent c5157fcad5
commit be0db2d74f
24 files changed
+1271 -96

No files matched your search

+1 -1
View File
@@ -29,7 +29,7 @@ The Prove tab reports every obligation in three statuses (`src/engines/pseudo/vc
| Shown | Earned by | Means | Does not mean |
|---|---|---|---|
| **Proved** | The built-in prover (`solver.ts`), on every path it examined, in a unit it examined completely | The claim follows from what is known, in the notation's number model: INTEGER as whole numbers of any size, FLOAT as finite numbers whose rounding is not seen into. The prover is sound (`tests/prover-soundness.test.mjs`) | That the Mojo made from the program has the same property. A 64-bit `Int` wraps and a `Float64` becomes `inf` or `nan`; the `int-model` and `float-model` notes say when a proof leans on the number model. A unit marked **partial** shows no claim as proved |
| **Proved** | The built-in prover (`solver.ts`), on every path it examined, in a unit it examined completely | The claim follows from what is known, in the notation's number model: INTEGER as whole numbers of any size, the sized types (INT8 … UINT64) exactly, with an `overflow` obligation wherever a value could leave its range, FLOAT and FLOAT32 as finite numbers whose rounding is not seen into. The prover is sound (`tests/prover-soundness.test.mjs`). In a program that says `USE IEEE` nothing is marked proved | That the Mojo made from the program has the same property. A 64-bit `Int` wraps and a `Float64` becomes `inf` or `nan`; the `int-model` and `float-model` notes say when a proof leans on the number model. A unit marked **partial** shows no claim as proved |
| **Broken** | A generated input (the fuzzer, seeded) or a run that breaks the claim at that place | There is an input for which the claim is false, and **Step through** shows it | Anything about other claims. A run outranks the prover: a proved obligation that an input breaks is shown as broken (and would be a soundness bug) |
| **Open** | Neither of the above | The checker could not finish the argument, usually because a loop needs an INVARIANT | That the claim is false |
| An invariant "the checker found on its own" | Inference in `vcgen.ts`, kept only at a fixed point | It holds on entry and survives every pass, and the proof used it | That it is the invariant a reader needs |